Justification Under Uncertainty: A Preliminary Inquiry
Posted: 16 May 2010
Date Written: May 14, 2010
People are sometimes asked to justify their decisions to courts, regulators, self-regulatory bodies, their colleagues or clients, or the “court of public opinion.” One might think the potential need to justify ex post would lead to better decisions ex ante. After all the better the decision ex ante, the less likely one will need to justify it ex post. But where considerable uncertainty exists as to the result of a decision, the potential need for justification may take center stage when the decision is being made. The existence of considerable uncertainty may mean that there is no straightforward way to make the best decision. Absent a straightforward way to make the best decision, people may instead focus on how they would justify the decision if there is a bad result. For some types of decisions, an early focus on justification may improve, or at least not appreciably impair, the decision-making. The decision may be made with more care than would otherwise have been the case, albeit perhaps at a cost higher than the commensurate benefit of the added increment of care. But in some cases, the potential need to justify may worsen the decision. I discuss three examples. One involves justification by using an accepted reference group: a money manager decides what to do based largely on what other reputable money managers are doing, or a CEO is selected from a pool of previous CEOs (perhaps including some who have been spectacular failures in past CEO positions). Another involves the use of a responsibility-shifting process such as delegation to a perhaps less qualified and knowledgeable person. A third involves the (over)use of “alignable metrics” – ways to rank order options- where it’s not clear that the metric is particularly relevant. If a firm does not have much experience with diverse employees, it might feel particularly uncertain regarding their future performance, and might rely too heavily on alignable metrics such as standardized test scores.
This phenomenon matters for law and policy makers: many decisions guided by justificatory concerns have serious social consequences. Massive overinvestment in subprime securities is an obvious example. Decisions to hire a CEO and the general trajectory of diversity hiring are also of societal import. For contexts in which a decision may be challenged in court, law of course encourages justification – law is part of the problem. And there is of course no easy solution. In these contexts, there is by definition no straightforward method that should yield the best decision; the obvious inclination is to point to what others are doing, what numbers are saying, and so on. Thus, it is hard to imagine law doing anything dramatically different than what it presently does. Indeed, law is far better at policing the thoroughness of process than its quality. Perhaps, though, dicta and norms can be marshaled to encourage better decisions. People presumably would rather make the best decision possible; justification is only a second best if the decision doesn’t work out. Perhaps decision-makers will focus more on the first best, a genuine best effort to make the best decision possible, if the divergence between the first and second best is made more salient.
Suggested Citation: Suggested Citation